search

reTHINK-project / core-framework

The main goal of WP3 is to provide the reTHINK core framework comprised by the runtime environment where Hyperties are executed and the messaging nodes used to support messages exchange between Hyperties.
Apache License 2.0
1 stars 0 forks source link

Runtime Security Analysis: complete and clean #75

Closed pchainho closed 9 years ago

pchainho commented 9 years ago

https://github.com/reTHINK-project/core-framework/blob/master/docs/specs/runtime/securityanalysis.md

aoncorici commented 9 years ago

completed, updated pictures, closing issue

pchainho commented 9 years ago

Reopening since this is an issue for Nuno

aoncorici commented 9 years ago

Sorry, I thought it was assigned to me.

nuno-santos commented 9 years ago

I had finished my revision before Antutás's pass. Since then, I've just noticed that a misspelling error was introduced in one of the figures: "Hiperty". Also, I'd ask Paulo to comment on Antutás's suggestion to change the title of last section to "M2M standalone application platform". I'm ok with that proposal.

pchainho commented 9 years ago

I would still prefer "Constrained Runtime". On the other hand I've noticed that architecture related figures were not corrected yet and are not aligned with text description. I propose the following changes:

Hyperty Sandbox -> ASP Sandbox Core Runtime -> Native Runtime

aoncorici commented 9 years ago

if we use the name "constraint runtime" we need to define why is it constraint and if we say energy we would get comments: but why do we consider that the mobile phone is not constraint (it also has battery, no endless energy resource). I would say that the constraint is that there is no human there. We could use: automation runtime, but it might be too much.

Let us change the name to be on the safe side when it comes to the outside review. Using M2M standalone application platform actually makes a point in the text provided by Nuno, that from the security point of view the platform is similar to the standalone platform, because it is one.

Anyhow we need a section defining the platforms (why we have 3 instead of 1, e.g.) and how the whole code will work together without replacing functionality. I think this is still missing from D3.1 and hope Paulo has an idea on the text to be included and act as an umbrella.

nuno-santos commented 9 years ago

Ancuta may have a point. Paulo, how would you like to proceed?

pchainho commented 9 years ago

Ok, however the current proposal for the standalone app is based on web views not on nodejs and sandbox implementation is quite different.

@nuno-santos have you merged / commented the pull request from Kevin:

https://github.com/reTHINK-project/core-framework/pull/91 ?

nuno-santos commented 9 years ago

Hi Paulo, no, didn't know I had to do that. What exactly do I have to do?

pchainho commented 9 years ago

Just click at #91 add you'll find there a place to comment and/or merge the changes made by Kevin

nuno-santos commented 9 years ago

Paulo, I'm confused by the latest version of the architecture diagram. In this new version, app and hyperty run in the same sandbox... Do you confirm this change? If so, the text from the security analysis is no longer consistent with the figure. There won't be any isolation between app and hyperty anymore.

pchainho commented 9 years ago

Nuno, as explained in the text and discussed in Aveiro, this should be possible in case we have more relaxed security policies eg when both are coming from the same domain. However, in case they don't trust each other you may have each one running in different sandboxes as depicted in the second picture:

unstrusted app and hyperty

nuno-santos commented 9 years ago

Ok, tried to address all aforementioned issues: (1) merged Kevin's contributions, (2) fixed the figures, (3) revised the terminology in the text to be consistent with the figure, (4) retained the title change of "constrained platform" to "m2m standalone" proposed by Ancuta, and (5) did another pass on the text. Hope everything is ok, now. So I'll close this issue.